Fig. 4

Tyrosine kinase inhibitors target ABCG2 function. A Representative images of ABCG2 fluorescent substrate Pheophorbide A upon treatment with KO143 (1 µM), lapatinib (3.16 µM), nilotinib (1 µM), rabusertib (1 µM) and THZ531 (0.1 µM) in Hs578T cells. Scale bar is 100 µm. B Relative pheophorbide A intensity upon treatment with inhibitors in Hs578T (upper) or SKBR7 (below) cells. Data are the normalized mean (± SD) of three independent experiments, means were compared using one-way ANOVA with Tukey’s correction for multiple testing. C Association of pheophorbide A accumulation with the amount of synergy THZ531 (difference in proliferation (percentage point) in combination treatment compared to additive effects of single treatments) of these kinase inhibitors with THZ531. Kinase inhibitors were screened at 1 µM and were classified (See methods) as having “No effect” (n = 112), “Weak/Inconsistent” (n = 83) or inducing “Pheophorbide A accumulation” (n = 103). Kinase inhibitors that already strongly affected proliferation as individual treatment (proliferation DMSO < 40%) were excluded, and synergy was determined by the difference between individual effects of THZ531 and KI’s. Means were compared using Kruskal–Wallis test with Dunn’s correction. D Spearman correlation of amount of synergy of kinase inhibitors with THZ531 (0.1 µM) and the normalized pheophorbide A accumulation upon treatment with these kinase inhibitors alone. E Association of predicted ABCG2 inhibitory activity of kinase inhibitors with the amount of synergy of these kinase inhibitors with THZ531. Kinase inhibitors were scored as inactive (n = 109), uncertain (n = 182), or active (n = 31) towards ABCG2 depending on the inhibitory score from the modelling. Kinase inhibitors that already strongly affected proliferation as individual treatment (proliferation DMSO < 40%) were excluded. Means were compared using Kruskal–Wallis test with Dunn’s correction